This is a prospective, non-randomized study of discontinuing prophylaxis in HIV-infected subjects who have had a CD4 count < 100 cells/mm3 any time in the past and no confirmed PCP, or who have had a confirmed episode of PCP more than 6 months prior to study entry. Subjects will discontinue their PCP prophylaxis if antiretroviral therapy has resulted in a sustained CD4 increase to > 200 cells/mm3 on two measurements at least 12 weeks apart. The primary objective is assess the incidence of PCP after discontinuation of PCP prophylaxis within 3 months of entry for these subjects. The secondary objectives are to assess the incidence of presumed PCP, HIV-related infections and diarrhea in subjects who have discontinued PCP prophylaxis within 3 months of study entry in the above stratified groups. Another objective is to compare the incidence of confirmed PCP for subjects who have not reinstituted prophylaxis with those who have reinstituted prophylaxis for PCP. The study will attempt to determine whether there is a correlation between changes in HIV/RNA or CD4 count and the incidence of PCP.